Ellenálló mobiléletciklus

Mobile Lifecycle & Resilience

Mobile apps run in an interruptible environment: phone calls, app switching, OS memory pressure, process death and radio handoffs (Wi-Fi to cellular) happen mid-operation. Design every flow to survive interruption without losing user work or leaking resources.

1. Handle every lifecycle transition (active / inactive / background / process death)

• On inactive / background (AppState in React Native, AppLifecycleListener in Flutter, ScenePhase / LifecycleObserver in native):
  • Immediately pause camera previews, audio/video playback, high-frequency timers, sensor/GPS polling and continuous animations.
  • Flush or checkpoint unsaved form drafts and in-progress state to local storage synchronously or immediately—the OS may kill the backgrounded process without further warning.
• On active (foreground resume):
  • Re-validate session/token expiry, refresh stale time-sensitive queries, and reconcile the status of any in-flight mutation or payment that was interrupted.
  • Re-check runtime permissions (camera, location, notifications) that the user may have toggled in OS Settings while away.
• OS process death & state restoration: never rely solely on in-memory singletons surviving backgrounding; restore screen routes by ID and reload data or persisted drafts on cold/warm restart.

2. Flaky networks, radio switches and interrupted transfers

• Connectivity != reachability: NetInfo / connectivity_plus reporting Wi-Fi does not mean the API is reachable (captive portals, blackholed cellular handoffs). Enforce explicit connect/read timeouts (10-30s) and distinguish network/timeout errors from 4xx business rejections.
• Never destroy user input on failure: if a form save, comment or upload fails due to a timeout or connection drop, keep the user's draft intact on screen and provide a clear Retry action.
• Background & chunked transfers: iOS and Android throttle background sockets within seconds; for large media uploads/downloads, show progress, support cancellation/resume, and use OS background transfer APIs (URLSession background config, Android WorkManager, expo-file-system background tasks) when work must survive app switching.

3. Single-flight token refresh and resource cleanup

• Guard 401 Unauthorized token refresh with a single-flight mutex/promise so 5 concurrent requests resuming after foregrounding await one refresh call instead of racing and triggering false logouts.
• Remove every AppState, NetInfo, keyboard, notification, stream and timer listener in useEffect cleanup / dispose() / onCleared().

Before you finish

• Backgrounding pauses sensors/media/timers and checkpoints unsaved drafts before potential OS process death.
• Foregrounding refreshes stale data, reconciles pending work and re-verifies OS permissions.
• Network timeouts/drops preserve user input and offer safe retry without infinite loops.
• Concurrent 401 token refreshes share a single-flight lock; all listeners and timers are cleaned up.